Background The efficacy of a levofloxacin-based regimen as the first-line treatment and a clarithromycin-based regimen as the second-line treatment for Helicobacter pylori infection remains unknown. The aim of this study was to assess the eradication rates of these two regimens using different administration sequences. Methods Eligible patients were randomised to receive LAL: levofloxacin (750 mg once a day), amoxicillin (1000 mg twice a day) and lansoprazole (30 mg twice a day) for 7 days, or CAL: clarithromycin (500 mg twice a day), amoxicillin (1000 mg twice a day) and lansoprazole (30 mg twice a day) for 7 days . Patients with positive [C- 13] urea breath test after treatment were retreated with the rescue regimen in a crossover manner for 10 days. Result When used as first-line treatment (n = 432), the eradication rates of LAL (n 217) and CAL (n 215) were 74.2 and 83.7% (p = 0.015) in the intent-to-treat (ITT) analysis, and 80.1 and 87.4% (p = 0. 046) in the per-protocol (PP) analysis, respectively. When used as second-line treatment, the eradication rates of LAL (n = 26) and CAL (n = 40) were 76.9 and 60% (p = 0.154) in the ITT analysis, and 80 and 61.5% (p = 0.120) in the PP analysis, respectively. The overall eradication rates of CAL followed by LAL were better than the reverse sequence in both the ITT analysis (93% vs 85.3%, p = 0.01) and the PP analysis (97.6% vs 92.5%, p = 0.019). The eradication rate was significantly lower in the presence of levofloxacin resistance in the LAL group (50% vs 84.4%, p = 0.018) and clarithromycin resistance in the CAL group (44.4% vs 90.7%, p = 0.002). Conclusion CAL achieved a higher eradication rate than LAL as the first-line treatment, but not as the second-line treatment. The strategy of using CAL as the initial treatment and LAL as the rescue regimen achieved higher eradication rates than the reverse sequence. ; 附設醫院內科部 ; 醫學院附設醫院 ; 期刊論文